Search code examples
c#linux.net-coreimagesharp

System.Drawing.Common.Bitmap cross platform alternatives

I am looking to convert PDF files into images. Docnet is able to convert the pdf into bytes[] and their samples show how to save this byte[] into an image file using Bitmap. Documentation

However, the solution won't work on linux machine since Bitmap requires few libraries pre-installed on the system.

I've tried ImageSharp to convert the byte[] using SixLabors.ImageSharp.Image.Load<Bgra32>(rawBytes), however, it throws Unhandled exception. SixLabors.ImageSharp.InvalidImageContentException: PNG Image does not contain a data chunk.

Does anyone knows any alternative to achieve this.

PS - I'm open to explore any other cross platform FREE supported alternatives to convert PDF files to images.

Solution

  • This works fine with ImageSharp assuming Docnet works then ImageSharp will work fine for you.

    The trick is you want to be using the Image.LoadPixelData<Bgra32>(rawBytes, width, height); API not the Image.Load<Bgra32>(encodedBytes); one.

    using Docnet.Core;
using Docnet.Core.Models;
using SixLabors.ImageSharp;
using SixLabors.ImageSharp.PixelFormats;
using SixLabors.ImageSharp.Processing;

using var docReader = DocLib.Instance.GetDocReader(
           "wikipedia_0.pdf",
           new PageDimensions(1080, 1920));

using var pageReader = docReader.GetPageReader(0);

var rawBytes = pageReader.GetImage();

var width = pageReader.GetPageWidth();
var height = pageReader.GetPageHeight();

// this is the important line, here you are taking a byte array that 
// represents the pixels directly where as Image.Load<Bgra32>()
// is expected an encoded image in png, jpeg etc format
using var img = Image.LoadPixelData<Bgra32>(rawBytes, width, height);
// you are likely going to want this as well otherwise you might end up with transparent parts.
img.Mutate(x => x.BackgroundColor(Color.White));

img.Save("wikipedia_0.png");